Date: Tue, 30 Nov 2010 00:44:50 -1000
From: Brendan Fox <jindeni@gmail.com>
Subject: Re: Local Testbed/ARA Meetings - Today (Monday, November 29th) @ 1:30PM Hawaii Time (6:30PM EST) in the IDLab

Summary from memory at 1AM by bullet points ...

1 - batwing global trigger efficiency is pretty bad

    I'll have Ben send you the talk with the plots in them, but the
batwing global trigger looks
    very bad right now because the noise tail extends out a long way
and thus requires us to
    use such a high threshold that we are only about 50% efficient at
an SNR of around 7-8.
    We think that this could be a result of an issue in the firmware
since, as the discriminator
    scalers update every 10 ms, we cannot see rates below 100 Hz in
them.  We're going to
    change the firmware to see those scaler rates this week.  It is
also possible that we have
    some board noise and that is given us a rate of correlated noise
triggers that is higher
    than we'd like.

2 - the counter reset by the Ru clock and stopped by each event does
not seem to work

     We're not sure what is wrong here as the counter looks okay when
tested with a short
     delay on the ara2 board.  But, with the real system, it is not
good.  We are planning to
     ChipScope this issue when we take the system out of the freezer
later this week.  At
     that time, we will also check to see if the input for the GPS
trigger is still alive since we
     might have killed that in the past.

3 - the 'event type' data in the housekeeping data needs to be
latched differently

     The event type, pattern, etc. housekeeping data is currently
latched every 1 ms, but
     really should be latched every event to make sure that there is
no 'stale' data problem.
     Luca and Patrick are going to change the readout of these data
so that the internal
     register, as opposed to the latched register, is readout by the
DAQ so as to ensure the
     data for a given event is not stale.

4 - taking calibration data now in the freezer

     Rishi is working on the zero-crossing calibration method using
data that was taken at
     room temperature.  I'll ask him to circulate some plots to you. 
The odd thing that we see
     in that the hitbus start/end position is not uniformly
distributed -- it is about 2x as likely to
     occur at the beginning or end of the capacitor line as it is to
occur someone in the middle.
     That strikes us as strange.

     Over the next two or so days, I plan to take data at different
sine wave amplitudes in addition
     to various frequencies; hopefully, I'll get a complete set of
data at two different temperatures.

5 - SBC memory problem

    The heater improves things, but not enough to avoid the problem. 
I am going to implement
    a 'sense that the SBC is down and power cycle the supply' script
on the central computer
    so that the system will stay up.  Peter is hopeful that we will
get some new memory from
    the Swiss and that'll cure our problem.

Brendan